# ForgeMesh Speech-to-Text

> ForgeMesh Speech-to-Text is a paid API for AI agents from x402.forgemesh.io, paid per call via x402, $0.03/call, status unknown (last checked 2026-09-14).

Transcribes audio files from a public URL into text using a neural speech recognition engine supporting ~99 languages and common audio formats.

## Facts

- Endpoint: POST https://x402.forgemesh.io/speech-to-text
- Price: $0.03/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-14
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/forgemesh-speech-to-text-6a538725
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_7wLM5FQzyrwD19wFfZbiB

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability forgemesh-speech-to-text-6a538725 -d '<json body>'
```

Example prompt: Transcribe this voicemail audio file for me — it's an mp3 at https://storage.example.com/voicemail-2024-06-10.mp3 and I need the full text of what was said.

## When to prefer this

Use this endpoint when you have a publicly accessible audio file URL and need a text transcript quickly without managing your own speech recognition infrastructure. Well-suited for voicemail pipelines, podcast transcription, meeting notes, and voice-agent workflows where privacy matters since audio is deleted post-processing. Supports ~99 languages and common formats (mp3, wav, m4a, ogg) up to 25MB per file.

## Known failure modes

- Audio file URL is not publicly accessible or returns 4xx/5xx — endpoint cannot fetch the file
- Audio file exceeds 25MB size limit — request rejected
- Unsupported audio format (not mp3, wav, m4a, or ogg) — processing fails
- Audio contains no intelligible speech — empty or low-quality transcript returned
- Network timeout fetching large audio files — processing error

## How this service works

Speech-to-text API: audio URL in, transcript out — a neural speech-recognition engine on our own hardware, ~99 languages, mp3/wav/m4a/ogg up to 25MB. For voicemail, podcast, meeting, and voice-agent pipelines. Audio deleted after processing.

## Output

A text transcript of the spoken content in the audio file, derived from neural speech recognition. The audio is deleted after processing. Supports approximately 99 languages.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "audio_url": {
   "type": "string",
   "description": "Public URL of an audio file, max 25MB"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "text": "The quick brown fox jumps over the lazy dog. ForgeMesh utility grid speech fixture."
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/forgemesh-speech-to-text-6a538725/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from x402.forgemesh.io](https://www.zero.xyz/host/x402.forgemesh.io/llms.txt)